This article was reprinted with permission from Superpower.com.
As the memory chip prices rise, Samsung Semiconductor's business has been a big harvest, and now has replaced Intel's 25-year-old status as the world's largest semiconductor process, and in advanced technology, Intel also can not hold the first position, 10nm and 7nm The craftsmanship was taken by TSMC and Samsung robbed it.
Even AMD has to release 7nm Vega chips this year. Intel's 10nm process is still difficult to produce. It will not be mass-produced in the second half of next year. The new process is difficult to produce. What about Intel? No need to do it, Intel CEO Kozange has already stated his position. They will also continue to optimize the 14nm process.
Since the publication of financial statements some time ago, the issue of difficult 10nm process production at Intel has become a key issue. Markets and analysts are concerned about this issue, and Intel can no longer find a good solution. The yield rate will not directly affect the production of processors. The output and cost of the process, the process encountered problems can not be solved in a flash, so at the earnings conference Intel CEO Kozanch also had to explain.
Intel stated that while they are working hard to solve the 10nm yield problem, they will continue to optimize the 14nm process. Kozanch emphasized that since the introduction of the 14nm process, they have improved their performance by 70% through process optimization. This ratio is not. Knowing how to count, Intel earlier announced that the third-generation 14nm process 14 nm + + only increased performance by 26% compared to the initial generation process.
Kozanci said they will continue to use 14nm to launch leading products, and later this year will launch the Whiskey Lake processor for the client market and the Cascade Lake processor for the data center.
On the other hand, Intel is emboldened to say that their 14nm process specifications are better than those of TSMC and Samsung's 16 and 14nm processes, and that both transistor density and transistor gate pitch are far ahead of their opponents. However, the current problem is that Intel In the 14nm process, the opponent has already upgraded to the 10nm and 7nm processes. The gap has been opened up to two generations, and the opponents have not been able to survive. The two generations of upgrades have also shortened or even countered the gap.
